This break index marks a medium degree of disjuncture between adjacent words. The boundary marked by BI 2 is stronger than that marked by BI 1, but it lacks the cues (e.g. lengthening, pauses, etc.) common to an even stronger boundary marked by BI 3 (see section 4.4).
In most cases the unit marked by a BI 2 at its boundary is characterized by the H- and L% delimiting tones of the accentual phrase, the level of the prosodic hierarchy above the word. However, this perceptually-defined unit (BI 2) and the tonally-defined unit (accentual phrase) are not always identical. There may occasionally be cases of mismatch between the perceived juncture and the tonal characteristics (this is described in detail in section 4.5 below).
The utterance 
sankaku
contains a good
example of break index 2. The words /sa'Nkaku/ `triangle' and /no/
`GEN' are grouped together into one tonally-defined unit (accentual
phrase), and the following words /ya'ne/ `roof' and /no/ `GEN' into
another. The perceived separation between words within each phrase (BI
1) is smaller than the separation between words belonging to adjacent
phrases (BI 2). The rise to the H- of the second accentual phrase
gives the sense of the beginning of a new unit, and thus we perceive
the boundary between the two units as stronger than that within units,
or stronger than if there was no phrase break present. In this
utterance, the unit marked by BI 2 is identical to the tonally-defined
accentual phrase.
Example utterances 
yane
and
narabu
also show break index 2 marking a
medium degree of disjuncture, which in these cases also is identical
to an accentual phrase break.